Massive Changes to My Engine Design ! 1

[RodRico]

All,

You may recall the challenges I was facing with installing a fuel injector in every cylinder of my engine; with a 49.5cc engine comprised of 6 cylinders, each was only 8.25cc and required a micro-fuel injector that had good enough spray pattern to yield a well mixed homogeneous fuel/air mixture. After working through the injector design, I decided the injector approach was simply too risky for a small engine such as my prototype. I thus went off to work up an alternate architecture using a single larger fuel injector to feed an intake manifold shared between all cylinders. This puts the fuel injector into the realm of components I can buy off-the-shelf.

The use of a shared intake manifold with premixed fuel/air unfortunately requires use of two air pump pistons in place of the one in the prior architecture; one air pump piston for scavenge fed with air alone and another for intake fed with mixed fuel and air. The use of separate pistons increases complexity because each must now be driven by independent cam lobes (not a big deal) but the two pistons don't affect performance much because the pistons are very light and operate at comparatively low pressure. They do, of course add some volume, but not as much as I expected. Since I was reworking everything, I decided to exploit the separate cams on intake and scavenge to mechanize over expansion (to increase efficiency) and simplify the means by which I was compensating for cold-start and altitude (which both require variable compression ratio). All of these changes are evident in the annotated illustration below.

While I'm bummed at the amount of work this change is taking to update math models, CAD models, CFD, and FAE, I'm happy with the performance and glad to have caught the problems in the prior architecture *before* I started cutting metal!

Comments, suggestions, and questions are encouraged!

Rod

 

[gruntguru]

A p-v diagram would be nice.

How are you calculating thermal efficiency?

2:1 expansion ratio with 60% TE???

je suis charlie

 

[RodRico]

The estimates are from my course thermodynamic model. Performance is estimated by calculating isentropic relations for start/end volume, pressure, and temperature for each of the major cycles that produce or consume significant work (pump compression, charge compression, combustion/expansion) at sea level. Gamma (y, ratio of specific heats) is calculated for each step based on temperature. Work (W) is then calculated for each step then summed and IMEP calculated. FMEP is the average of Heywood and Ricardo figures for typical engines. BMEP is then calculated and used to calculate HP, Torque, efficiency, and BSFC. Next, I reduce efficiency by 8.7% to reflect typical heat loss in an opposed piston engine then recalculate performance at the lower efficiency. I don't bother calculating air pump intake or the scavenge cycle because both are low pressure processes using low tension rings and should have little impact on total work (assuming properly sized ports).

I'm now updating the more detailed model that uses actual cam timing and Hohenberg's heat transfer coefficient to calculate heat loss for each one-half degree of the cycle. This model is a lot more work because it incorporates a host of detailed characteristics (port area, cam timing, reciprocating mass, cam loads, etc.), so I never undertake it until the preliminary model looks good (meaning it shows excess performance relative to goals).

I misspoke regarding expansion ratio (the ratio of P3|P4 volume at exhaust vs volume at intake). Though I *set* it at 2:1, the preliminary model limits expansion to ensure minimum conditions when the exhaust port first opens (the lesser of 4 bar pressure or 850F temperature). As it stands, the compression ratio of the air pump (P1|P2) is 1.44:1 while the main cylinder (P3|P4) has a compression ratio of 18:1 and an expansion volume equal to 1.16x the intake volume before compression starts.

 

[gruntguru]

So an expansion ratio of 18 x 1.16 = 20.88:1?

That makes a bit more sense.

je suis charlie

 

[enginesrus]

I'm waiting for the big change. And of course the running prototype. The main cam drive system won't last or is it now a crankshaft?

 

[RodRico]

gruntguru, It seems I've abused terms again. Yes, the ratio of volumes from combustion to opening of the exhaust valve is 20.88:1.

enginesrus, I spent months struggling to explore the micro-fuel injectors, and I'll spend months reworking all the details of the new arrangement having 5 pistons per cylinder set in place of the prior 3, so I think it's a pretty big change. A working prototype is my only objective, but it's still going to be a while; a new engine design based on existing and well proven concepts typically takes a *team* of engineers a couple of years, so my progress is obviously going to be markedly slower. I acknowledge I may well fail for any number of reasons, but I don't know of any that are assured as you imply in your cam comment. The mass, travel, and acceleration of the pistons in the prototype engine are well below those associated with the valve train in a larger conventional engine, so I don't expect premature failure of the cams to be an issue. The cams may become a critical issue at some point in larger engines, but I'm not planning to build anything larger than about 650cc and am open to using multiple rotors, so the cams may never be an issue. Only time, prototypes, and endurance testing will tell.

Rod

 

[enginesrus]

I guess its my opinion, but the cam as used in the normal reciprocating engines that we all know, is a weak point especially in racing applications, and other than the exhaust stroke it is not loaded as much as what a system like yours will be, it is just not a good way to convert reciprocating motion to rotary motion, and actually there is much to be considered in cam design such as accelerations and such. You would be better off using the technique that some hydraulic motors do.

 

[BigClive]

If a conventional diesel engine had an expansion ratio of 20.88:1 (to the point when the exhaust valve opened) it would mean that the geometrical compression/expansion ratio would be about (at a guess) 40:1. You get the same effect with a conventional petrol/ICE engine - if the CR is 9:1 and the temps of combustion and exhaust are typically about 1900/700 degrees C - this equates to about an actual effective 4:1 expansion ratio. Sorry about the guesses but I have worked it out more accurately before and I recall that the figures are something like these.

I still haven't seen a general simple diagram of how the engine operates (but I may have missed it).

 

[RodRico]

BigClive,

I may have changed my target exhaust temp since I wrote the post, but I don't think I've changed it much. Here are the key parameters...

Air Pump
Start Volume: 2.76E-05 m3
Start Pressure: 101,493 Pa
Start Temp: 288K
Air Mass: 3.39E-05 kg (note some will be lost in manifold during transfer to compression cylinder)
y: 1.380 (gamma, ratio of specific heats)
End Volume: 2.25E-05 m3
End Pressure: 134,175 Pa
End Temp: 311K

Compression
Start Volume: 1.01E-05 m3
Start Pressure: 134,175 Pa
Start Temp: 311K
Air Mass: 1.52E-05 kg (mass in 12.375 cc volume at sea level pressure and temp)
y: 1.332 (gamma, ratio of specific heats)
End Volume: 4.88E-07 m3
End Pressure: 7,613,756 Pa
End Temp: 852K

Combustion and Expansion
Start Volume: 4.88E-7 m3
Start Pressure: 21,000,000 Pa (210 bar, a design max)
Start Temp: 2,000K (peak combust temp, a design max)
Air mass: 1.52E-05 kg
y: 1.280 (gamma, ratio of specific heats)
End Volume: 1.35E-05 m3
End Pressure: 298,979 Pa
End Temp: 788K (chosen to ensure catalytic converter operation and at least 2 bar delta to ambient pressure).

Cylinder Performance
Work: 11.35 J per each of four cylinders (includes expected 17.4% heat loss)
IMEP: 9.17 bar per each of four cylinders
FMEP: 0.53 bar per each of four cylinders (average of Heywood and Ricardo)
BMEP: 8.65 bar

Engine Performance
Displacement: 49.5 cc in all (four) cylinders combined (12.375 cc each)
HP: 10.1 all cylinders combined
Torque: 20.1 lb-ft all cylinders combined
Equivalency: 0.403 (36.4:1 air to fuel ratio)
Efficiency: 60%
BSFC: 0.230 lb/hr/HP

Fuel Characteristics
Type: Diesel #2
LHV: 42.8 MJ/kg
Stoichiometric AFR: 14.5:1
Ignition Delay at 311K: 5.2 ms
Ignition Delay at 852K: 1.0 ms​

From the above, the Air Pump compression ratio is 1.22:1 and the Cylinder compression ratio is 20.72 yielding a combined 27.7:1 compression ratio while the expansion ratio is 27.7:1. I'm not sure if, when discussing expansion ratio, it should be compared to Air Pump * Cylinder compression ratio or just Cylinder compression ratio. It's really only a matter of terms, however, and the safest way is to provide all three figures as I have. Regardless of wording, the result is calculated as the air moves through the engine stages and I'm reasonably confident my calculations are correct. It's very important in reviewing these figures to note the engine uses HCCI combustion, the closest approximation to constant volume combustion, which does not spray fuel during the expansion stroke.

Rod

P.S. The illustration I posted at the top of this thread was intended to be a simplified representation of how the engine works. I presented it as though the cylinders were stationary while cams turned to move pistons to avoid the confusion I usually encounter when trying to explain operation when cylinders are rotating and cams are stationary as they are in the *actual* design (a Rotating-Cylinder Radial Two-Stroke Opposed-Piston Miller-Cycle HCCI Engine ).

 

[SwinnyGG]

I'm not sure if, when discussing expansion ratio, it should be compared to Air Pump * Cylinder compression ratio or just Cylinder compression ratio.

Expansion ratio describes the mechanical behavior during the portion of the cycle during which power is extracted.

Unless the 'air pump piston' is generating power, the expansion ratio is just the reciprocal of the compression ratio of the actual working swept volume, ie 20.7.

 

[RodRico]

jgKRI,

In that case, I am "over expanded" meaning the ratio of (volume at exhaust port opening)/(volume at ignition) which is 27.7:1 is greater than the compression ratio of the primary cylinder alone which is 20.7:1. This difference is the result of my using a "Miller Cycle" in the main working cylinder... I hold the intake port open while the piston moves up a bit from the exhaust port so the compression stroke is shorter than the expansion stroke.

Rod

 

[SwinnyGG]

I feel like we've been down this road before (arguing about the fine details of terminology) and at this point I have a great deal of respect for the work you're trying to do, but man... this is why people get confused when you try to explain your concept to them.

In your diagram above, the P3/P4 pair are exposed to combustion pressure, while P1/P2/P5 are used only for managing airflow, and are never exposed to combustion pressure- correct? When you say 'air pump piston', which piston are you referring to?

If that is the case, the expansion ratio is the ratio between maxima and minima of that swept volume.

For your engine it's more complicated than for a simpler conventional piston-rod-crank architecture, sure- but the expansion ratio, as a property, does not care about the physical linkage used to control the change in volume. You have the added wrinkle of making it possible to have compression and expansion ratios which are, in fact, different. But expansion ratio still represents only the ratio of the two instantaneous states.

Is the ratio of maximum to minimum volume of the P3/P4 area 27.7? If so, I agree then that's what you should call 'expansion ratio'.

 

[RodRico]

JgKRI,

P1/P2 are the “Intake Air Pump,” P3/P4 are the “Compression/Expansion” pistons, and P5 is the “Scavenge Air Pump.” The P3/P4 pistons have a 20.7:1 compression ratio and a 27.7:1 expansion ratio.

I know I stumble over terminology. It’s caused by the fact that I don’t speak to experienced engine designers anywhere but here (and occasionally with a consultant) and the engine is inherently confusing because it’s so different. The math is unambiguous, and I’ll eventually get the descriptive terms right.

The thermodynamics don’t care how I effect compression, and that’s why I got confused as to whether the Intake Air Pump compression ratio should be included in overall compression ratio.

Rod

 

[RodRico]

I feel this could soon be the perfect shirt for me!

 

[michaelwoodcoc]

If you're interested in working together i was working on a tiny injector design. Really it's been thought of before. I'm just planning subtle changes. Thibk if a linear compressor in a fridge. Your injector could also be the fuel pump. The biggest thing for me is trying to figure out the electronics. Fridges have a lit of space for inverters and electronics. If theres2 no small solution it negates removal of fuel pumo

 

[RodRico]

michaelwoodcoc,

One of the challenges in the fuel injector (besides the terrifically small volume it must inject) is the fact that it resides inside a rotating cylinder block. The injector I was designing was thus a unit type cam driven mechanical injector... one cam lobe (fixed timing) created the pressure applied to the needle while another (variable timing) released that pressure to shorten the inject event and thus reduce injected volume. I still think it's a good approach, it just seemed to be dominating my risk... everything was very small with tight tolerances. My objective is to prove the basic engine, not develop a new injector, so I started evaluating other options. If you have something that eliminates the time and risk of the micro-injector, I'd love to hear about it.

Rod

 

[michaelwoodcoc]

ok, I understand more throughly now. I'm not sure it's possible for your situation what I had in mind. I was consiering a few different ways but I cannot find a simple way of putting it on a rotating cylinder block. It could be used as kindof a port injection I suppose.

 

[RodRico]

michaelwoodcoc,

OK, since the in-rotor micro-injector is off the table for now, we can discuss the manifold injector related to the reworked design described in this post.

The injector for the design described in my post is mounted in a stationary manifold, so it can be any reasonable size and either mechanically or electrically controlled. At present, I'm considering two options:

1) Off the shelf injector pump ($150) and injector ($100) from a 10 HP Carroll Stream diesel. This approach would require I add a cam to drive the injector pump.

2) Ultrasonic Atomizer(s) using a high output transducer ($20) driven by a programmable controller ($445). Note the controller cost will come way down once settings are determined and the programmable controller replaced with a fixed controller based on a reference design.

Though it's less certain, I prefer #2 because it should provide smaller droplet size with less penetration depth than the injector, and its installation would be much simpler (electronics versus adding a cam to drive an injector pump). The uncertainty can be eliminated via experiments and a $500 investment, so it's not a terrible risk.

Rod

 

[michaelwoodcoc]

It's kindof difficult to think about this without a clear picture in my head. I think you're really going for very well atomized fuel given your consideration for an ultrasonic transducer and a high pressure diesel injector setup, am I correct to think that?

Port diesel injection is something i have not considered before. Are you going to be running the engine at rapidly varyibg loads? We have 49cc engines in mopeds with injectors that may be to small. Perhaps a zuma 125 injector would work.

I haven't done any calculations or looked at viscosity charts or anything like that but you may consider preheating diesel with waste engine heat can get it very close or to the flash point, just so long as that is far away from the i jector melting point. I've seen fuel preheating on biodiesel, diesel, and even regular gasoline applications.

it's infinitesimal, but directly injecting cold fuel in a diesel would sap some energy from expanding the air. With port fuel injection I'm nsure the fuel should be able to soak up the needed heat on it's way to the chamber.

 

[RodRico]

Yes, the "Homogeneous Charge" part of "Homogeneous Charge Compression Ignition" (HCCI) explicitly requires a well mixed charge going into compression. Because the fuel/air mixture would pass from the stationary side housings to the spinning rotor, I could add fins to create turbulence and improve mixing, but that would take Joules, so it's better to have a well mixed charge before the transfer.

I'm not sure how you quantify "rapidly varying loads," but I assume you're alluding to the fact that manifold injection will be sluggish compared to direct injection into the cylinder. The market for my engine is light aviation, so production engines would make between 76 lb-ft torque at 2,626 RPM (RQ-7 drone) and 230 lb-ft torque at 2,626 RPM (Rotax 914 UL) at 15,000 feet. I don't think these type aviation engines need to be highly responsive and, if they do, they may be large enough to make the in-cylinder injectors feasible. Note I *may* sell the 49.5 cc prototype for use in scooters, lawn equipment, and model aircraft if there's a demand, but I don't think the long term outlook in such markets is attractive as they are rapidly transitioning to electric (as they should IMHO).

I have always planned on heating fuel (and the block) electrically during cold start at -40F, so I assume electric block and fuel heaters. Once the engine is running at equilibrium, exhaust heat can be used to aid the injectors. For reference, exhaust temp is currently predicted to be about 634F once the engine is warmed up, and I will create no design with exhaust temps below 550F as required by the typical catalytic converter.